Compounding for maximum heat resistance and load bearing capacity in HNBR belts.

Hydrogenated nitrile-butadiene rubber (HNBR) continues to experience significant growth in many industrial and automotive applications due to its combination of high temperature, aggressive fluid and oil resistance. However, the major market for HNBR remains in automotive belts.

Environmental concerns over global warming (from so-called greenhouse gases such as C[O.sub.2]) have led to a number of technology changes in the automotive sector aimed at improved fuel efficiency and reduced hydrocarbon emissions. These include:

* Autobody styling (lower drag coefficients and increased cabin space);

* vehicle weight reduction (SMC body panels, etc.);

* more fuel efficient |lean burn' engines (reduced hydrocarbon emission);

* wider use of catalytic converters (Europe).

In general, these changes limit the engine size and weight as well as reducing the size of the engine bay area and cooling radiator. In order to maintain acceptable power output, automobile manufacturers are making extensive use of multivalve engines (MVEs) which allow higher specific output for a given size compared to a traditional two-valve engine. The peak power and torque for the MVEs comes at 10-15% higher revolutions, thus, to improve lower end output the manufacturers are turning to computer controlled variable valve timing and, in some cases, also variable valve lift (e.g. Honda's VTEC system).

Desirable properties

Considering the above, then, it can be seen that the demands on the synchronous belt have increased significantly. The vehicle manufacturers, having already moved from polychloroprene to more heat-resistant HNBR belts, now require the following improvements:

* Durability increase of 1.5 times existing HNBR belt;

* 20% narrower package (i.e. narrower belt width);

* 10 [degrees] C increased heat resistance.

The synchronous belt is a composite unit made from an elastomeric compound, fiber tension member (glass or aramid) and a polyamide cross-woven facing fabric.

The belt's function of transmitting power to the camshaft must be done with the minimum of dynamic deformation in order to prevent |time lag' between the pulley and the belt. In addition, the deformation under a given force should remain independent across the operating temperature and rpm range of the engine. These needs are even more essential in the new variable timing/lift MVEs.

Because of its location, the belt may encounter oil spray. The compound should, therefore, resist swelling from engine oils to prevent a reduction in the dynamic stiffness of the teeth. As a composite of rubber compound and fiber, adhesion between the components is vital.

The work reported in this article describes the effect of cure system, carbon black and alternative fillers on the load bearing capacity (modulus) and heat resistance (air) of HNBR synchronous belt compounds. The work is designed to show how the elastomeric HNBR compound can be formulated to produce a significant increase in hot air resistance with higher modulus (load bearing capacity) while maintaining other desirable characteristics essential to the performance of the synchronous belt (oil resistance, stable dynamic modulus).

Experimental

Compounds for determining the effects of filler level and temperature on the dynamic modulus of HNBR, and filler loading on the dynamic modulus stability, were mixed in a Brabender with a rotor speed of 50 rpm and a chamber temperature of 100 [degrees] C. All other compounds were mixed in a laboratory "00" internal mixer using a five minute cycle, 77 rpm rotor speed and an initial rotor temperature of 50 [degrees] C. Curatives were added to the stock on a two-roll mill. The mill was run using 25 rpm in front and 35 rpm in the rear (1:1.4) and the initial roll surface temperature was 50 [degrees] C.

Various characteristics were determined using the following test methods:

* Viscosity and scorch characteristics - ASTM D 1646.

* Cure characteristics - using an MDR 2000E with a 0.5 [degrees] rotor arc at 160 [degrees] C.

* Stress-strain results - ASTM D 412 using a die C dumbbell.

* Hardness - ASTM D 2240.

Dynamic test results were obtained from a variety of testing equipment:

* Isochronal (11 Hz) temperature sweeps at very low strain amplitude ([+ or -]0.025%) were made in extension on an Autovibron.

* Low strain amplitude single point bulk moduli were determined using a Mechanical Energy Resolver 1100A.

* Heat build-up and dynamic set were determined using a Goodrich Flexometer operating from 115 [degrees] C with 17.5% deflection and 10.9 kg load.

Cure system effects

Current technology HNBR belts use thiuram-based cure systems which result principally in disulphidic crosslinks. Sulfur and thiuram based curing systems utilize unsaturation in their crosslinking mechanism. Logically, reducing the unsaturation should lead to improved heat resistance. However, this also reduces the potential crosslink sites leading to lower maximum modulus as shown in figure 1.

The bond energies of linkages possible from thiuram and peroxide cures are shown in table 1.

Table 1.
Type of X-link   Bond energy
                  (kcal/mol)
  S-S                64
  C-S                68
  C-C                84

Energetically the C-C bond is favored and it is well established that the heat stability of peroxide-cured compounds is superior to sulfur or thiuram types. Therefore, to improve the heat resistance of the belt compound without an undesirable reduction of the modulus, the use of a peroxide cure system is preferable.

Effect of carbon black

Mechanisms

Before discussing the effect of carbon black in HNBR synchronous belt compounds it is useful to review some general effects of carbon black reinforcement.

Reinforcement of elastomers by carbon black fillers is considered to occur through a number of physical and chemical mechanisms including, but not exclusive to:

* Van der Waal's forces;

* hydrogen abstraction from the aromatic nuclei of the carbon black (ref. 1);

* adsorption of the polymer chain onto the carbon black surface.

Modulus increase in black-filled elastomers results from hydrodynamic interactions with an effect known in general as strain amplification. This latter term encompasses work from A.M. Bueche (ref. 2) who originally suggested an expression to account for the inextensible carbon black particles, a modification by F. Bueche (ref. 3) to include the fact that occluded rubber could not deform in the same manner as the gum network and, specifically, Mullins and Tobin (ref. 4) who reasoned that the local strain in the gum phase of a filled system increased by a factor X over the macroscopic strain of the sample. Kraus (ref. 5) developed a carbon black structure-concentration equivalence principle for strain amplification effects due to carbon black structure.

Strain amplitude dependence of black filled compounds under dynamic deformation is well established from the classic work of Payne (ref. 6). In this work it was shown that increasing amounts of filler produce an ever higher amplitude dependence at low strain. Payne suggested an interaggregate model to account for the observed phenomenon. In a recent review together with experimental work, Funt (ref. 7) suggested an alternate model based on effective crosslink density due to bulk rubber entanglements with filler-bound rubber.

Practical application

While the most obvious approach to increased load bearing capacity is to increase die carbon black content or use a more reinforcing type (smaller particle size, higher structure), in view of the above mechanistic assessment it is clear that such practice would have some negative consequences to the earlier stated performance criteria of a synchronous belt compound. Although the strain amplification effects of carbon black reinforcement are desirable, the amplitude dependency is not (irrespective of which model is correct).

The effect of temperature and filler concentration on dynamic modulus (E*) is shown in figure 2 for N550 black in a peroxide-cured HNBR with 2% residual double bonds (Tornac B 3850). As can be seen, the dynamic modulus increase is both dependent upon filler loading and temperature.

Isolating the responses of 25 [degrees] C and 150 [degrees] C the temperature-loading interrelation can be more clearly seen (figure 3). As the filler loading increases beyond 30 phr the increase in dynamic modulus follows two separate third-order polynomials.

Recently, Pouchelon and Vondracek (ref. 8) discussed a percolation model as a basis for a power law description of mechanical property changes with filler content for various elastomers and fillers. The work showed that the shear elastic modulus increased according to a power-law beyond the loading of filler needed to generate filler-filler network. This filler loading was termed the threshold value. Using electron microscopy, all filler concentrations below this value (or point of percolation), were shown to have isolated particles or were agglomerated in finite isolated clusters. The filler level at the percolation point varies for different fillers and elastomers dependent upon the nature of the filler and the filler-elastomer interaction. Data in figure 2 would suggest compliance with the semi-empirical approach of Pouchelon and Vondracek.

Figure 4 shows a series of carbon black types from ASTM N100 to ASTM N700 series, compounded at 50 phr in a sulfur-cured HNBR of 6% residual double bonds. Two additional compounds with 30 and 70 phr N650 black are also included. N234, N347, N339, N351, N550 and N650 all have approximately the same DBP value (125).

The high dynamic modulus of the N100 (SAF), N200 (ISAF) and N300 (HAF) series carbon blacks exhibited at room temperature diminishes at 150 [degrees] C. While the modulus of the N550 and N650 are lower than the reinforcing blacks at room temperature, they show similar modulus at 150 [degrees] C. Thus, it is seen that the expected benefits of more reinforcing carbon blacks are not realized for this application.

Goodrich heat build-up measurements for the same series of carbon blacks are shown in figure 5. The effect of filler loading is clearly seen for the three N650 filler levels (30, 50 and 70 phr).

The effect of strain amplitude can be seen in figure 6. It is important to note that the least change in modulus is seen for the low level of carbon black of large particle size.

From the above it is clear, that in order to reduce temperature and amplitude dependence, the amount of filler should be kept below the percolation point to prevent filler-filler network perturbations. This leads to a significant drawback with respect to the earlier stated narrower package performance goal set down by the automotive companies. In order to provide the additional load bearing capacity further reinforcement is required that is derived from a polymer bound agent. This need can be met using a modification of existing rubber technology from the rubber-roll cover and golf ball industries.

Resin reinforcement

For many years methacrylate coagents have been used in NBR and EPDM to produce high hardness roll compounds. A U.S. patent granted to Fukahori and Kataoka (ref. 9) in 1981 describes the reinforcement of NBR by an in situ formed zinc dimethacrylate (from zinc oxide and methacrylic acid). The patent specifically noted the need for the presence of a paraphenylene diamine (PPD) antiozonant in order to achieve maximum reinforcement. It is possible that the PPD acts to prevent the combination of elastomer radicals thus encouraging homopolymerization of the zinc dimethacrylate. Eventual termination of the poly (zinc dimethacrylate) to the polymer would result in a rubber-bound polymeric filler. HNBR has very little unsaturation and requires twice as much peroxide curative than NBR to obtain the same crosslink density. Therefore, not surprisingly, the PPD is not required for HNBR. In situ formed zinc dimethacrylate has also been used in the golf ball industry but, due to corrosion effects of the acid on the mixing equipment, this practice has essentially ceased in favor of pre-formed salts. A comparison of zinc dimethacrylate and zinc diacrylate (table 2) of the same average particle size indicates that the diacrylate salt would produce higher modulus compounds for the same weight % loading.

Table 2 - comparison of zinc salts
Formulation (phr)
  HNBR (<0.5% RDB)                        100    100
  Zinc di-methyacrylate (ZDMA)             35
  Zinc di-acrylate (ZDA)                          35
  [alpha], [alpha]'bis (t-butlperoxy)-      5      5
  -diisopropylbenzene [40% active]
Stress/strain
  Tensile strength (MPa)                 20.7   27.7
  Modulus at 100% elong. (MPa)            1.8    4.5
  Modulus at 200% elong. (MPa)            4.3    9.9
  Modulus at 300% elong. (MPa)            8.1   18.9
  Ultimate elongation (%)                 550    385

Comparison of heat-resistant belt compounds

In this evaluation a comparison is made between thiuram cured HNBR and a mixed black/zinc diacrylate reinforced peroxide-cured compound utilizing the merits of the previous discussion. Also included for comparison is an ACSM (an alkylated chlorosulphonated polyethylene) compound. ACSM is a recently introduced candidate for heat-resistant belts. The formulations are shown in table 3. The zinc diacrylate is incorporated into an HNBR (5% RDB) based masterbatch. The masterbatch consists of the zinc salt, cure retarders and viscosity reduction agents of a proprietary nature. The total rubber hydrocarbon is 50% by weight. The ACSM formulation was taken from recent work by Pillow and Ennis (ref. 10).

Table 3 - formulations for heat-resistant belt
compounds

                                       phr   phr   phr
HNBR (5% RDB)[Therban 1707S HRD]       100    -     75
ACSM (27% Cl)                           -    100    -
Acrylate zinc salt masterbatch          -     -     50
N326                                    -     -     25
N762                                    -     35    -
N650                                    45    -     -
6PPD                                   1.0    -     -
TMQ                                    1.0    -    1.0
ZMMBI                                  1.0    -    1.0
Zinc oxide                             5.0    -   10.0
Stearic acid                           1.0    -     -
Tri-octyl trimellitate                 7.0    -    5.0
NBC                                     -    3.0    -
Paraffin wax                            -    2.0    -
Low M.Wt. polyethylene                  -    3.0    -
Pentaerythritol                         -    3.0    -
Litharge (100%)                         -   10.8    -
HVA-2                                   -    1.0    -
Magnesium oxide                         -    4.0    -
DPTT                                    -    1.0    -
Sulphur                                0.5    -     -
TMTD                                   1.5    -     -
MBTS                                   1.5   1.0    -
TMTM                                   0.5    -     -
[alpha], [alpha]'bis (t-butlperoxy)-    -     -    8.0
diisopropylbenzene [40% active]

Table 4 shows the cure characteristics and basic stress/strain data for the candidate compounds. The sulfur cured HNBR and ACSM compounds were vulcanized 30 minutes at 160 [degrees] C while the peroxide-cured HNBR was vulcanized for 45 minutes at 160 [degrees] C. The peroxide cured HNBR and the ACSM formula show similar hardnesses and moduli but the HNBR shows typically high tensile strength and higher ultimate elongation. [TABULAR DATA OMITTED]

The failure modes of a synchronous belt are usually associated with, but not exclusive to:

* Cracking of the back cover due to loss of extension on heat aging;

* breakage of teeth due to loss of strength after heat aging;

* jacket failure due to permanent set of the teeth compound (or swelling from oil).

The effects of exposure to hot air, for various periods of time, on the tensile strength and ultimate elongation are shown in figures 7 and 8. Up to 336 hours the ACSM and sulphur-cured HNBR compounds both perform well. At 504 hours the elongation of the ACSM is less than 50% and, in addition, the tensile strength has also dropped to a very low level while in the sulfur-cured HNBR, tensile strength remains reasonably high.

The tensile strength of the peroxide-cured HNBR is relatively unaffected after 504 hours and still retains 50% of its original elongation.

Both HNBR candidates show low permanent set and excellent resistance to swelling from oil. On the other hand, the permanent set of ACSM is much worse and its resistance to oil swell is poor. Increasing the chlorine content of the ACSM polymer will improve its resistance to oil at the expense of low temperature properties (ref. 10).

Low temperature performance of belt compounds can be measured by brittleness point and a flexibility test which involves wrapping a strip of compound around a mandrel following a conditioning period at low temperature (24 hours at -40 [degrees] C in this case). Results for the candidate compounds are shown in table 5. [TABULAR DATA OMITTED]

The failure of the ACSM compound is most likely due to the poor integral strength after aging (see figure 7).

Figure 8 depicts the isochronal temperature response of the three compounds from 0 [degrees] C to 150 [degrees] C and shows that the dynamic modulus of the peroxide-cured compound is considerably higher than either the sulfur-cured HNBR or ACSM compounds. It is interesting to note that the higher static modulus of the ACSM compound (from table 3) does not translate into high dynamic modulus, indicating considerable stress softening (Mullin's effect).

As indicated earlier, a synchronous belt is a composite of fabrics and compound. Adhesion results shown in table 6 indicate that equivalent adhesion can be obtained for either sulfur or peroxide cured HNBR with existing glass fiber cord prepared for sulphur cured HNBR.

Table 6 - adhesion to glass fiber cord

                          HNBR              HNBR
                     (sulfur cured)   (peroxide cured)
Maximum value (kg)        44                 42
Minimum value (kg)        38                 38

Summary and conclusions

The demands of the synchronous belt continue to increase with respect to heat resistance, load bearing capacity and longevity. A review of factors affecting heat resistance and modulus stability was made. Finally, this article has described the basic compounding principles which, when applied to HNBR, can be used to successfully achieve significant property enhancements towards meeting the new demands of engine designers without detracting from the desirable properties of a stable dynamic modulus and adhesion to fabric components. Peroxide cured HNBR, particularly when reinforced with a mixture of carbon black and zinc diacrylate, shows excellent promise to meet the emerging needs of synchronous drive belts.

References

1. E. Papirer, A. Voet and P.H. Given, "Transfer of labeled hydrogen between elastomers and carbon black," Rubber Chem. Technol. 52,1200 (1964). 2. A.M. Bueche, J. Polym. Sci., 25, 139 (1957) - via J.A.C. Harwood, L. Mullins and A. Payne, "Stress softening in rubbers - a review," J. of the IRI, 17, Jan/Feb., 1967. 3. F. Bueche, J. Polym. Sci., 5,271 (1961) ibid. 4. L. Mullins and N.R. Tobin, J. Polym. Sci., 9, 2993 (1965) ibid. 5. G. Kraus, Rubber Chem. Technol., 44, 199 (1971). 6. A. Payne, J. Applied Polym. Sci., 7, 873 (1963). 7. J.M. Funt "Dynamic testing and reinforcement of rubber," ACS paper 54, May 1987. 8. A. Pouchelon and P. Vondracek, "Semi-empirical relationships between properties and loading in filled elastomers," Rubber Chem. Technol. 62, 788 (1989). 9. Y Fukahori and N. Kataoka, U.S. patent 4,332,918, June 1, 1982. 10. J. Pillow and R. Ennis, "New generation of chlorinated elastomer," "ACS paper 19, Oct. 1989.